Plur-e MCP · Model Context Protocol
Talk to Business Central and Dynamics 365 CRM — with AI
Plur-e MCP servers expose your ERP and CRM as governed tools for any MCP client: real data, real actions, your users' permissions, full audit.
- Example prompts
- “What's on hand for item PROD-001 across all locations?”
- “Create a sales quote for Customer example 1 with 40 cases of PROD-001 at list price.”
- “Which purchase orders are late this week, and who is the vendor contact?”
- “Summarize the last five cases for Account example 1 and draft a follow-up.”
Quick answer
Plur-e MCP is a set of Model Context Protocol (MCP) servers for Microsoft Dynamics 365 Business Central and Dynamics 365 CRM (Dataverse). They let your AI assistant — connected through any MCP client (desktop, IDE or web) or custom agents built on your AI provider's API — read and, when allowed, write ERP and CRM data through tools with per-user authentication, MCP roles (Read / Write / Post) and full logging. Plur-e designs, deploys and operates them for customers, in Plur-e's cloud or in the customer's Azure tenant.
How it works
An AI client connects
Any MCP client — desktop, IDE or web — or your own agent built on your AI provider's API or agent SDK.
Plur-e MCP exposes governed tools
Each tool has a name, a schema and a scope. Read tools are on by default; write and post tools are enabled per MCP role.
Calls run as the user
Authentication is per user with Microsoft Entra ID (OAuth). Your AI assistant never sees more than the user's MCP role and the ERP identity mapped to them allow.
Everything is logged
Every request, tool call and result is recorded. Writes carry the acting user, the payload and the resulting document number.
Tools
MCP for Business Central
Warehouse, orders, items, sales, payments and time sheets — the 262 operations of Plur-e Mobile as 16 section tools, plus barcodes, capture and printing.
ToolsMCP for Dynamics 365 CRM
Any Dataverse table, Customer Insights - Journeys campaigns and per-tenant memory.
ToolsPlur-e Docs MCP
Public, read-only server over this documentation. Add Plur-e docs to your AI assistant.
ToolsGovernance first
| Identity | Per-user OAuth with Microsoft Entra ID; no shared service account for user-facing tools |
|---|---|
| Default posture | Read-only. Write and post tools are opt-in per MCP role |
| Scopes | Tools grouped by area (sales, purchasing, inventory, warehouse, CRM) and enabled per customer and role |
| Audit | Full request and tool-call log with user, timestamp, inputs and outputs |
| Rate limits | Per user and per tenant, configurable |
| Data | Prompts and tool results are not used to train models; retention follows your AI provider's commercial terms. |
| Hosting | Plur-e cloud (Azure) or your Azure tenant |
Why Plur-e for MCP
- We already run Business Central integrations in production: Plur-e Mobile posts warehouse, sales and payment documents every day. The same operations are now AI tools.
- We know the Business Central and Dataverse APIs, permissions and posting rules — the parts that make an agent safe, not just impressive.
- We build in .NET and TypeScript and ship MCP servers with tests, logs and versioned tool schemas.
- We have shipped custom MCP tools tailored to customers' Business Central and CRM workflows.
- We operate an AI assistant inside Business Central (the Chat Bot Assistant).
Looking for the assistant inside Business Central? That is the Chat Bot Assistant. AI Assistant
Your AI assistant can also trigger the same receiving, picking and shipping flows your team runs in Plur-e Mobile. Plur-e Mobile
How an engagement runs
How an engagement runs
Discovery workshop (2 h)
Map the questions your teams ask every day, the data behind them and who may act on what.
Pilot (2–4 weeks)
Three to five tools for one team, connected to a sandbox or production with read-only scopes. Real users, real prompts.
Production
Governance (roles, audit logs), monitoring, training and documentation. Optional write tools.
Managed service
We operate the servers, ship new tools as your processes change and keep up with Business Central releases.
Metered usage: pay per tool call your AI assistant makes, plus an optional one-time onboarding fee. Request a quote.
FAQ
Frequently asked questions
Start with a two-hour discovery workshop
Bring the questions your team asks every day. We map them to tools, permissions and a pilot plan.